文摘
Variational methods in imaging are nowadays developing towards a quite universal and flexible tool, allowing for highly successful approaches on tasks like image restoration, registration, segmentation, super-resolution, and estimation of flow fields. We review recent progress in mathematical image processing by combining first and second order derivatives in the regularization term of variational models. We demonstrate the power of the proposed methods by two rather different applications. The approaches make use of two different splitting methods of the functional to obtain iterative numerical schemes which require in each step only the computation of simple proximal mappings.